Abundance analysis for long period variables

Velocity effects studied with O-rich dynamic model atmospheres

¹ University of Vienna, Department of Astronomy, Türkenschanzstraße 17, 1180 Wien, Austria e-mail: thomas.lebzelter@univie.ac.at
² Department of Physics and Astronomy, Division of Astronomy and Space Physics, Uppsala University, Box 515, 75120 Uppsala, Sweden
³ National Optical Astronomy Observatories (Operated by the Association of Universities for Research in Astronomy, Inc. under cooperative agreement with the National Science Foundation.) , 950 N. Cherry Avenue, Tucson, Arizona 85726, USA
⁴ Osservatorio Astronomico di Padova – INAF, vicolo dell'Osservatorio 5, 35122 Padova, Italy

Received: 24 November 2009
Accepted: 31 March 2010

Abstract

Context. Measuring the surface abundances of AGB stars is an important tool for studying the effects of nucleosynthesis and mixing in the interior of low- to intermediate mass stars during their final evolutionary phases. The atmospheres of AGB stars can be strongly affected by stellar pulsation and the development of a stellar wind, though, and the abundance determination of these objects should therefore be based on dynamic model atmospheres.

Aims. We investigate the effects of stellar pulsation and mass loss on the appearance of selected spectral features (line profiles, line intensities) and on the derived elemental abundances by performing a systematic comparison of hydrostatic and dynamic model atmospheres.

Methods. High-resolution synthetic spectra in the near infrared range were calculated based on two dynamic model atmospheres (at various phases during the pulsation cycle) as well as a grid of hydrostatic COMARCS models with effective temperatures T_eff and surface gravities log g over an adequate range. Equivalent widths of a selection of atomic and molecular lines (Fe, OH, CO) were derived in both cases and compared with each other.

Results. In the case of the dynamic models, the equivalent widths of all investigated features vary over the pulsation cycle. A consistent reproduction of the derived variations with a set of hydrostatic models is not possible, but several individual phases and spectral features can be reproduced well with the help of specific hydrostatic atmospheric models. In addition, we show that the variations in equivalent width that we found on the basis of the adopted state-of-the-art dynamic model atmospheres agree qualitatively with observational results for the Mira R Cas over its light cycle.

Conclusions. The findings of our modelling form a starting point to deal with the problem of abundance determination in strongly dynamic AGB stars (i.e., long-period variables). Our results illustrate that some quantities such as the C/O ratio can probably still be determined to a reasonable accuracy, but the measurement of other quantities will be hampered by the dynamics. The qualitative agreement with observations of R Cas opens promising possibilities for a forthcoming quantitative comparison of our synthetic spectra with observed ones of AGB variables in the globular cluster 47 Tuc.